A Model Predictive Control Strategy for Overtaking in Autonomous Vehicles

This paper proposes a model predictive control (MPC) strategy for planning and dynamic obstacle avoidance in autonomous vehicles. The problem consists of two vehicles with different speeds sharing the same road. The objective is to safely allow the faster vehicle to overtake the slower one. The proposed strategy integrates path planning and control into a single optimization problem. Moreover, it formulates nonlinear constraints for lateral deviation and collision avoidance and linearizes them into half spaces to tackle the optimization problem more easily using linear inequality constraints. The proposed approach ensures a safe overtaking maneuver by utilizing information about the predicted positions of the slower vehicle over the MPC prediction horizon of the faster one. Simulation results demonstrate that the proposed MPC strategy can successfully execute safe and smooth overtaking maneuvers on roads with various speeds and curvature.